The Savanna Elephant Loxodonta africana (Blumenbach 1779) evolved with the other live-bearing mammals (Theria) from the Cynodonts (mammal-like reptiles) of the Triassic (225-195 million years ago). First to diverge from the stem were egg-laying mammals (Monotremes) during the Jurassic age. In the early Cretaceous era (about 130 million years ago), the Theria diverged into 3 major groups, one of which was the Marsupials. The other two became the placental mammals (Eutheria).

The earliest ungulates, the Condylarthra, appeared at the end of the Cretaceous period some 65 million years ago and the Subungulates evolved from an offshoot of this group early in the Paleocene. The modern ungulates are now placed in an entirely separate group (Laurasiatheria).

The group which included the ancestors of modern elephants (Afrotheria) diverged further during the Paleocene and separated into several distinct orders. The Aardvark (Orycteropus), regarded by some taxonomists as belonging to the Subungulata, diverged early in the Paleocene, preceded by the Tenrecs, Golden Moles and Xenarthrans.

By the early Eocene (54 million years ago), the Subungulates proper consisted of 3 distinct orders - the Dugongs and Manatees (Sirenia), the Hyraxes (Hyracoidea) and the elephant progenitors.

The first proboscidean was Phosphatherium (58 million years ago) from which Moeritherium evolved at the start of the Oligocene. Amongst the numerous elephant 'models' which followed were Gomphotherium, Trilophodon, and Platybelodon and, in the Pleistocene, the Imperial Mammoth which became extinct in recent times. The taxonomy of Macdonald (2001) has been used in this study.

Although no subspecies of the Savanna elephant are recognised today, this has not always been the case. Shortridge (1934) recognised two subspecies: L.a. knochenhauri and, of relevance here, L.a. zukowskyi from the type locality of Qoabendus in the Kaokoveld (Figure 3). Shortridge notes that this is "a large elephant" but discredits Steinhardt's claim that it reaches 4.5 metres at the shoulder. He also notes that the elephants that inhabit the Kaokoveld and south-western Angola appear to be isolated from other populations and that Wilhelm considered this subpopulation to differ from the Caprivi- Okavango elephants. Shortridge speculates that South-eastern Angola, the Caprivi, Ngamiland and north-western Zimbabwe may in fact be a single geographical area for elephants - which corresponds closely with Parker's (1979) findings. Large tusks were not common amongst these elephants.